Daylight-Sensitive Efficient Lighting

In just a decade the light-emitting diode (LED) has moved from being the favored option for indicators on electronic equipment to the primary focus in making lighting more energy efficient. The ability to generate intense light levels efficiently has made the LED an increasingly popular choice for most applications, surpassing the abilities of fluorescent and other technologies. Another important advantage of the LED over other options is its controllability.

Although dimming an LED involves a more complex circuit design than the simple controllers used for traditional incandescent lighting, the LED supports the ability to tune not just the lighting level, but the emitted color to a fine level of adjustment. This controllability can be exploited in a variety of applications, from high-bay lighting in warehouses and factories to smart greenhouses.

LED lighting is already contributing to increased yield in agriculture by extending the possible growing season for a wide range of crops by several months in temperate climates. By improving yield and growing times, the technology makes rooftop agriculture more economic. As a result, cities can begin to contribute to global food production. An important aspect of smart greenhouse design is the ability to make the LEDs respond to ambient light levels to avoid overconsumption of electricity and to provide different colors to support the needs of crops over their lifecycle. The LEDs can be turned off completely during bright ambient lighting conditions, and be gradually brought online as it decreases.

In warehouses ambient lighting is less readily available, but it is often present through skylights and windows. The problem with traditional lighting design is providing consistent illumination over the entire floor space. Traditionally, luminaires will be set to provide a constant level of illumination that will be wasteful in areas of high ambient light levels. At the same time, areas in corners with few surrounding luminaires and windows or skylights may be comparatively dark. Smart LED lighting can compensate for these problems by adapting not only to changing natural light, but the impact of other luminaires.

A smart LED lighting solution combines circuitry to control each element with sensors that detect ambient light levels. This combination provides the basis for efficient light harvesting without demanding complex communication schemes, although an increasing number of systems are incorporating wireless microcontrollers to support interactive control.

In principle, the light provided by an LED is easy to control. The output of an LED is an approximately linear function of forward current, and to provide consistent lighting, it requires a constant drive current. Typically, the brightness versus current curve flattens out as the LED junction heats up, wasting energy due to the characteristic loss of efficiency with an increase in temperature.

Figure 1: Graph of LED output versus current.

In practice, the current flow to the LED can be interrupted for short periods as long as the frequency is faster than can be detected by the human eye. In general, a switching frequency of 200 Hz avoids situations where people notice the lights are flickering. The effective current and therefore the perceived brightness level of the LED become proportional to the duty-cycle of the waveform provided by the dimming circuitry. This makes pulse-width modulation (PWM) an effective way of achieving efficient LED dimming.

An LED power controller designed for PWM operation such as the Texas Instruments TPS92640 provides a dimming range of 2500:1 using standard PWM dimming, and 20000:1 with the aid of a shunt FET. The use of a shunt FET on another member of the family, the TPS92641, overcomes the restrictions that the minimum PWM cycle time imposes on dimming levels. Because it takes time to charge and discharge the output stage inductor of a PWM power controller, more current than is required for a given level of light may need to be passed through the LED. By activating a shunt FET that is wired in parallel to the LED or series-wired string of LEDs, much of the current can be diverted. The TPS92640 and similar controllers provide a connection to the shunt FET to activate for a variable portion of the cycle and therefore dim the LED further.

A further advantage of PWM dimming is that it provides better control over the LED’s output spectrum, an important aspect of white LEDs that use phosphor coatings. At a low current the phosphor output dominates, which tends to result in a yellowish hue. At a high current the blue light of the LED dominates. By driving the LED with a specific current, the circuit designers can assure a consistent color output. This is not possible using a constant current strategy where the current level itself needs to be altered to achieve different dimming levels.

The TPS92640 and TPS92641 accept either an analog or a PWM input from an external controller. A smart lighting system will use a microcontroller to implement the sensing logic and lighting control, which tends to favor the application of using a PWM input. A device such as the CC2650 microcontroller made by Texas Instruments provides the ability to connect to low-power wireless networks that operate at 2.4 GHz, and numerous interface ports to accept sensor signals.

Figure 2: Block diagram of a smart, light-harvesting luminaire.

The Texas Instruments OPT3001 can provide information to the system about existing light levels. The spectral response of the sensor was designed to be tightly matched to the photopic response of the human eye. Its filtering includes significant infrared rejection in order to prevent the possibility of overestimating the amount of harvestable ambient light.

Providing results over the I²C bus, measurements can be made by the OPT3001 from 0.01 lux up to 83 klux without the need to manually select full-scale ranges. This is performed using an internal full-scale setting feature. This capability allows light measurement over a 23-bit effective dynamic range. Data is extracted from each measurement word sent over the I²C by separating the most significant nibble. This upper nibble contains a scaling factor that needs to be applied to the three lower nibbles to provide the value in lux.

The OPT3001 can perform measurements continuously or as single-shot readings, integrating readings over a period of 100 ms or 800 ms to remove the effects of noise sources from other artificial light sources. The device has its own interrupt logic. This allows its use in a design where the host MCU can sleep until a limit has been passed. This improves efficiency in situations where the luminaire is not expected to be active for long periods – such as bright sunny days. The OPT3001 can compare its most recent reading with the high-limit and low-limit registers and generate an interrupt if either of those has been encountered. The MCU can reprogram the limit registers over I²C to support different modes of operation. For example, once the luminaire is active, the MCU can accept readings from the OPT3001 and does not need the sensor to interrupt based on the high or low limits until the next mode change. Alternatively, the MCU can shut down the OPT3001 to save power and use single-shot readings taken at regular intervals. After each single-shot reading, the OPT3001 shuts down.

Some system designs may choose to integrate the OPT3001 into the luminaire itself. Others may use distributed OPT3001 sensors coupled to wireless microcontrollers to relay light-intensity data from floor level back to the luminaires, with the local MCU inside each luminaire integrated the results to compute the PWM ratio required by the LED string itself.

Figure 3: Connections between the OPT3001 and a host MCU.

A further important component within the smart luminaire is a temperature sensor. This can be used to ensure that the LED is protected from over temperature conditions. The MCU can respond to these not just by reducing power supplied to the LED in order to protect the devices but to raise an alarm over the wireless network. A suitable temperature sensor is the Texas Instruments LMT84. This is a precision CMOS temperature sensor providing an analog output that can be used to feed into one of the CC2650’s ADC ports. The output voltage of the LMT84 is linearly and inversely proportional to temperature. The design has transient load requirements that are suited to drive a typical ADC sample-and-hold input and so provide continuous monitoring of thermal conditions within the luminaire to improve lifetime and efficiency.

Conclusion

Overall, by combining sensors with MCU-based intelligence and suitable power converters, it is possible to build smart luminaires that not only maximize the use of natural light, but which cooperate with each other to provide consistent illumination over large spaces.